mbed official / mbed-dsp

CMSIS DSP library

Dependents:   performance_timer Surfboard_ gps2rtty Capstone ... more

Legacy Warning

This is an mbed 2 library. To learn more about mbed OS 5, visit the docs.

cmsis_dsp/FilteringFunctions/arm_conv_partial_f32.c@2:da51fb522205, 2013-05-30 (annotated)

Committer:
emilmont
Date:
Thu May 30 17:10:11 2013 +0100
Revision:
2:da51fb522205
Parent:
1:fdd22bb7aa52
Child:
3:7a284390b0ce
Keep "cmsis-dsp" module in synch with its source

Who changed what in which revision?

UserRevisionLine numberNew contents of line
emilmont 1:fdd22bb7aa52 1 /* ----------------------------------------------------------------------------
emilmont 1:fdd22bb7aa52 2 * Copyright (C) 2010 ARM Limited. All rights reserved.
emilmont 1:fdd22bb7aa52 3 *
emilmont 1:fdd22bb7aa52 4 * $Date: 15. February 2012
emilmont 2:da51fb522205 5 * $Revision: V1.1.0
emilmont 1:fdd22bb7aa52 6 *
emilmont 2:da51fb522205 7 * Project: CMSIS DSP Library
emilmont 2:da51fb522205 8 * Title: arm_conv_partial_f32.c
emilmont 1:fdd22bb7aa52 9 *
emilmont 2:da51fb522205 10 * Description: Partial convolution of floating-point sequences.
emilmont 1:fdd22bb7aa52 11 *
emilmont 1:fdd22bb7aa52 12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
emilmont 1:fdd22bb7aa52 13 *
emilmont 1:fdd22bb7aa52 14 * Version 1.1.0 2012/02/15
emilmont 1:fdd22bb7aa52 15 * Updated with more optimizations, bug fixes and minor API changes.
emilmont 1:fdd22bb7aa52 16 *
emilmont 1:fdd22bb7aa52 17 * Version 1.0.11 2011/10/18
emilmont 1:fdd22bb7aa52 18 * Bug Fix in conv, correlation, partial convolution.
emilmont 1:fdd22bb7aa52 19 *
emilmont 1:fdd22bb7aa52 20 * Version 1.0.10 2011/7/15
emilmont 1:fdd22bb7aa52 21 * Big Endian support added and Merged M0 and M3/M4 Source code.
emilmont 1:fdd22bb7aa52 22 *
emilmont 1:fdd22bb7aa52 23 * Version 1.0.3 2010/11/29
emilmont 1:fdd22bb7aa52 24 * Re-organized the CMSIS folders and updated documentation.
emilmont 1:fdd22bb7aa52 25 *
emilmont 1:fdd22bb7aa52 26 * Version 1.0.2 2010/11/11
emilmont 1:fdd22bb7aa52 27 * Documentation updated.
emilmont 1:fdd22bb7aa52 28 *
emilmont 1:fdd22bb7aa52 29 * Version 1.0.1 2010/10/05
emilmont 1:fdd22bb7aa52 30 * Production release and review comments incorporated.
emilmont 1:fdd22bb7aa52 31 *
emilmont 1:fdd22bb7aa52 32 * Version 1.0.0 2010/09/20
emilmont 1:fdd22bb7aa52 33 * Production release and review comments incorporated
emilmont 1:fdd22bb7aa52 34 *
emilmont 1:fdd22bb7aa52 35 * Version 0.0.7 2010/06/10
emilmont 1:fdd22bb7aa52 36 * Misra-C changes done
emilmont 1:fdd22bb7aa52 37 *
emilmont 1:fdd22bb7aa52 38 * -------------------------------------------------------------------------- */
emilmont 1:fdd22bb7aa52 39
emilmont 1:fdd22bb7aa52 40 #include "arm_math.h"
emilmont 1:fdd22bb7aa52 41
emilmont 1:fdd22bb7aa52 42 /**
emilmont 1:fdd22bb7aa52 43 * @ingroup groupFilters
emilmont 1:fdd22bb7aa52 44 */
emilmont 1:fdd22bb7aa52 45
emilmont 1:fdd22bb7aa52 46 /**
emilmont 1:fdd22bb7aa52 47 * @defgroup PartialConv Partial Convolution
emilmont 1:fdd22bb7aa52 48 *
emilmont 1:fdd22bb7aa52 49 * Partial Convolution is equivalent to Convolution except that a subset of the output samples is generated.
emilmont 1:fdd22bb7aa52 50 * Each function has two additional arguments.
emilmont 1:fdd22bb7aa52 51 * <code>firstIndex</code> specifies the starting index of the subset of output samples.
emilmont 1:fdd22bb7aa52 52 * <code>numPoints</code> is the number of output samples to compute.
emilmont 1:fdd22bb7aa52 53 * The function computes the output in the range
emilmont 1:fdd22bb7aa52 54 * <code>[firstIndex, ..., firstIndex+numPoints-1]</code>.
emilmont 1:fdd22bb7aa52 55 * The output array <code>pDst</code> contains <code>numPoints</code> values.
emilmont 1:fdd22bb7aa52 56 *
emilmont 1:fdd22bb7aa52 57 * The allowable range of output indices is [0 srcALen+srcBLen-2].
emilmont 1:fdd22bb7aa52 58 * If the requested subset does not fall in this range then the functions return ARM_MATH_ARGUMENT_ERROR.
emilmont 1:fdd22bb7aa52 59 * Otherwise the functions return ARM_MATH_SUCCESS.
emilmont 1:fdd22bb7aa52 60 * \note Refer arm_conv_f32() for details on fixed point behavior.
emilmont 1:fdd22bb7aa52 61 *
emilmont 1:fdd22bb7aa52 62 *
emilmont 1:fdd22bb7aa52 63 * <b>Fast Versions</b>
emilmont 1:fdd22bb7aa52 64 *
emilmont 1:fdd22bb7aa52 65 * \par
emilmont 1:fdd22bb7aa52 66 * Fast versions are supported for Q31 and Q15 of partial convolution. Cycles for Fast versions are less compared to Q31 and Q15 of partial conv and the design requires
emilmont 1:fdd22bb7aa52 67 * the input signals should be scaled down to avoid intermediate overflows.
emilmont 1:fdd22bb7aa52 68 *
emilmont 1:fdd22bb7aa52 69 *
emilmont 1:fdd22bb7aa52 70 * <b>Opt Versions</b>
emilmont 1:fdd22bb7aa52 71 *
emilmont 1:fdd22bb7aa52 72 * \par
emilmont 1:fdd22bb7aa52 73 * Opt versions are supported for Q15 and Q7. Design uses internal scratch buffer for getting good optimisation.
emilmont 1:fdd22bb7aa52 74 * These versions are optimised in cycles and consumes more memory(Scratch memory) compared to Q15 and Q7 versions of partial convolution
emilmont 1:fdd22bb7aa52 75 */
emilmont 1:fdd22bb7aa52 76
emilmont 1:fdd22bb7aa52 77 /**
emilmont 1:fdd22bb7aa52 78 * @addtogroup PartialConv
emilmont 1:fdd22bb7aa52 79 * @{
emilmont 1:fdd22bb7aa52 80 */
emilmont 1:fdd22bb7aa52 81
emilmont 1:fdd22bb7aa52 82 /**
emilmont 1:fdd22bb7aa52 83 * @brief Partial convolution of floating-point sequences.
emilmont 1:fdd22bb7aa52 84 * @param[in] *pSrcA points to the first input sequence.
emilmont 1:fdd22bb7aa52 85 * @param[in] srcALen length of the first input sequence.
emilmont 1:fdd22bb7aa52 86 * @param[in] *pSrcB points to the second input sequence.
emilmont 1:fdd22bb7aa52 87 * @param[in] srcBLen length of the second input sequence.
emilmont 1:fdd22bb7aa52 88 * @param[out] *pDst points to the location where the output result is written.
emilmont 1:fdd22bb7aa52 89 * @param[in] firstIndex is the first output sample to start with.
emilmont 1:fdd22bb7aa52 90 * @param[in] numPoints is the number of output points to be computed.
emilmont 1:fdd22bb7aa52 91 * @return Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
emilmont 1:fdd22bb7aa52 92 */
emilmont 1:fdd22bb7aa52 93
emilmont 1:fdd22bb7aa52 94 arm_status arm_conv_partial_f32(
emilmont 1:fdd22bb7aa52 95 float32_t * pSrcA,
emilmont 1:fdd22bb7aa52 96 uint32_t srcALen,
emilmont 1:fdd22bb7aa52 97 float32_t * pSrcB,
emilmont 1:fdd22bb7aa52 98 uint32_t srcBLen,
emilmont 1:fdd22bb7aa52 99 float32_t * pDst,
emilmont 1:fdd22bb7aa52 100 uint32_t firstIndex,
emilmont 1:fdd22bb7aa52 101 uint32_t numPoints)
emilmont 1:fdd22bb7aa52 102 {
emilmont 1:fdd22bb7aa52 103
emilmont 1:fdd22bb7aa52 104
emilmont 1:fdd22bb7aa52 105 #ifndef ARM_MATH_CM0
emilmont 1:fdd22bb7aa52 106
emilmont 1:fdd22bb7aa52 107 /* Run the below code for Cortex-M4 and Cortex-M3 */
emilmont 1:fdd22bb7aa52 108
emilmont 1:fdd22bb7aa52 109 float32_t *pIn1 = pSrcA; /* inputA pointer */
emilmont 1:fdd22bb7aa52 110 float32_t *pIn2 = pSrcB; /* inputB pointer */
emilmont 1:fdd22bb7aa52 111 float32_t *pOut = pDst; /* output pointer */
emilmont 1:fdd22bb7aa52 112 float32_t *px; /* Intermediate inputA pointer */
emilmont 1:fdd22bb7aa52 113 float32_t *py; /* Intermediate inputB pointer */
emilmont 1:fdd22bb7aa52 114 float32_t *pSrc1, *pSrc2; /* Intermediate pointers */
emilmont 1:fdd22bb7aa52 115 float32_t sum, acc0, acc1, acc2, acc3; /* Accumulator */
emilmont 1:fdd22bb7aa52 116 float32_t x0, x1, x2, x3, c0; /* Temporary variables to hold state and coefficient values */
emilmont 1:fdd22bb7aa52 117 uint32_t j, k, count = 0u, blkCnt, check;
emilmont 1:fdd22bb7aa52 118 int32_t blockSize1, blockSize2, blockSize3; /* loop counters */
emilmont 1:fdd22bb7aa52 119 arm_status status; /* status of Partial convolution */
emilmont 1:fdd22bb7aa52 120
emilmont 1:fdd22bb7aa52 121
emilmont 1:fdd22bb7aa52 122 /* Check for range of output samples to be calculated */
emilmont 1:fdd22bb7aa52 123 if((firstIndex + numPoints) > ((srcALen + (srcBLen - 1u))))
emilmont 1:fdd22bb7aa52 124 {
emilmont 1:fdd22bb7aa52 125 /* Set status as ARM_MATH_ARGUMENT_ERROR */
emilmont 1:fdd22bb7aa52 126 status = ARM_MATH_ARGUMENT_ERROR;
emilmont 1:fdd22bb7aa52 127 }
emilmont 1:fdd22bb7aa52 128 else
emilmont 1:fdd22bb7aa52 129 {
emilmont 1:fdd22bb7aa52 130
emilmont 1:fdd22bb7aa52 131 /* The algorithm implementation is based on the lengths of the inputs. */
emilmont 1:fdd22bb7aa52 132 /* srcB is always made to slide across srcA. */
emilmont 1:fdd22bb7aa52 133 /* So srcBLen is always considered as shorter or equal to srcALen */
emilmont 1:fdd22bb7aa52 134 if(srcALen >= srcBLen)
emilmont 1:fdd22bb7aa52 135 {
emilmont 1:fdd22bb7aa52 136 /* Initialization of inputA pointer */
emilmont 1:fdd22bb7aa52 137 pIn1 = pSrcA;
emilmont 1:fdd22bb7aa52 138
emilmont 1:fdd22bb7aa52 139 /* Initialization of inputB pointer */
emilmont 1:fdd22bb7aa52 140 pIn2 = pSrcB;
emilmont 1:fdd22bb7aa52 141 }
emilmont 1:fdd22bb7aa52 142 else
emilmont 1:fdd22bb7aa52 143 {
emilmont 1:fdd22bb7aa52 144 /* Initialization of inputA pointer */
emilmont 1:fdd22bb7aa52 145 pIn1 = pSrcB;
emilmont 1:fdd22bb7aa52 146
emilmont 1:fdd22bb7aa52 147 /* Initialization of inputB pointer */
emilmont 1:fdd22bb7aa52 148 pIn2 = pSrcA;
emilmont 1:fdd22bb7aa52 149
emilmont 1:fdd22bb7aa52 150 /* srcBLen is always considered as shorter or equal to srcALen */
emilmont 1:fdd22bb7aa52 151 j = srcBLen;
emilmont 1:fdd22bb7aa52 152 srcBLen = srcALen;
emilmont 1:fdd22bb7aa52 153 srcALen = j;
emilmont 1:fdd22bb7aa52 154 }
emilmont 1:fdd22bb7aa52 155
emilmont 1:fdd22bb7aa52 156 /* Conditions to check which loopCounter holds
emilmont 1:fdd22bb7aa52 157 * the first and last indices of the output samples to be calculated. */
emilmont 1:fdd22bb7aa52 158 check = firstIndex + numPoints;
emilmont 1:fdd22bb7aa52 159 blockSize3 = (int32_t) check - (int32_t) srcALen;
emilmont 1:fdd22bb7aa52 160 blockSize3 = (blockSize3 > 0) ? blockSize3 : 0;
emilmont 1:fdd22bb7aa52 161 blockSize1 = ((int32_t) srcBLen - 1) - (int32_t) firstIndex;
emilmont 1:fdd22bb7aa52 162 blockSize1 = (blockSize1 > 0) ? ((check > (srcBLen - 1u)) ? blockSize1 :
emilmont 1:fdd22bb7aa52 163 (int32_t) numPoints) : 0;
emilmont 1:fdd22bb7aa52 164 blockSize2 = ((int32_t) check - blockSize3) -
emilmont 1:fdd22bb7aa52 165 (blockSize1 + (int32_t) firstIndex);
emilmont 1:fdd22bb7aa52 166 blockSize2 = (blockSize2 > 0) ? blockSize2 : 0;
emilmont 1:fdd22bb7aa52 167
emilmont 1:fdd22bb7aa52 168 /* conv(x,y) at n = x[n] * y[0] + x[n-1] * y[1] + x[n-2] * y[2] + ...+ x[n-N+1] * y[N -1] */
emilmont 1:fdd22bb7aa52 169 /* The function is internally
emilmont 1:fdd22bb7aa52 170 * divided into three stages according to the number of multiplications that has to be
emilmont 1:fdd22bb7aa52 171 * taken place between inputA samples and inputB samples. In the first stage of the
emilmont 1:fdd22bb7aa52 172 * algorithm, the multiplications increase by one for every iteration.
emilmont 1:fdd22bb7aa52 173 * In the second stage of the algorithm, srcBLen number of multiplications are done.
emilmont 1:fdd22bb7aa52 174 * In the third stage of the algorithm, the multiplications decrease by one
emilmont 1:fdd22bb7aa52 175 * for every iteration. */
emilmont 1:fdd22bb7aa52 176
emilmont 1:fdd22bb7aa52 177 /* Set the output pointer to point to the firstIndex
emilmont 1:fdd22bb7aa52 178 * of the output sample to be calculated. */
emilmont 1:fdd22bb7aa52 179 pOut = pDst + firstIndex;
emilmont 1:fdd22bb7aa52 180
emilmont 1:fdd22bb7aa52 181 /* --------------------------
emilmont 1:fdd22bb7aa52 182 * Initializations of stage1
emilmont 1:fdd22bb7aa52 183 * -------------------------*/
emilmont 1:fdd22bb7aa52 184
emilmont 1:fdd22bb7aa52 185 /* sum = x[0] * y[0]
emilmont 1:fdd22bb7aa52 186 * sum = x[0] * y[1] + x[1] * y[0]
emilmont 1:fdd22bb7aa52 187 * ....
emilmont 1:fdd22bb7aa52 188 * sum = x[0] * y[srcBlen - 1] + x[1] * y[srcBlen - 2] +...+ x[srcBLen - 1] * y[0]
emilmont 1:fdd22bb7aa52 189 */
emilmont 1:fdd22bb7aa52 190
emilmont 1:fdd22bb7aa52 191 /* In this stage the MAC operations are increased by 1 for every iteration.
emilmont 1:fdd22bb7aa52 192 The count variable holds the number of MAC operations performed.
emilmont 1:fdd22bb7aa52 193 Since the partial convolution starts from from firstIndex
emilmont 1:fdd22bb7aa52 194 Number of Macs to be performed is firstIndex + 1 */
emilmont 1:fdd22bb7aa52 195 count = 1u + firstIndex;
emilmont 1:fdd22bb7aa52 196
emilmont 1:fdd22bb7aa52 197 /* Working pointer of inputA */
emilmont 1:fdd22bb7aa52 198 px = pIn1;
emilmont 1:fdd22bb7aa52 199
emilmont 1:fdd22bb7aa52 200 /* Working pointer of inputB */
emilmont 1:fdd22bb7aa52 201 pSrc1 = pIn2 + firstIndex;
emilmont 1:fdd22bb7aa52 202 py = pSrc1;
emilmont 1:fdd22bb7aa52 203
emilmont 1:fdd22bb7aa52 204 /* ------------------------
emilmont 1:fdd22bb7aa52 205 * Stage1 process
emilmont 1:fdd22bb7aa52 206 * ----------------------*/
emilmont 1:fdd22bb7aa52 207
emilmont 1:fdd22bb7aa52 208 /* The first stage starts here */
emilmont 1:fdd22bb7aa52 209 while(blockSize1 > 0)
emilmont 1:fdd22bb7aa52 210 {
emilmont 1:fdd22bb7aa52 211 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 212 sum = 0.0f;
emilmont 1:fdd22bb7aa52 213
emilmont 1:fdd22bb7aa52 214 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 215 k = count >> 2u;
emilmont 1:fdd22bb7aa52 216
emilmont 1:fdd22bb7aa52 217 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 218 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 219 while(k > 0u)
emilmont 1:fdd22bb7aa52 220 {
emilmont 1:fdd22bb7aa52 221 /* x[0] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 222 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 223
emilmont 1:fdd22bb7aa52 224 /* x[1] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 225 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 226
emilmont 1:fdd22bb7aa52 227 /* x[2] * y[srcBLen - 3] */
emilmont 1:fdd22bb7aa52 228 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 229
emilmont 1:fdd22bb7aa52 230 /* x[3] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 231 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 232
emilmont 1:fdd22bb7aa52 233 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 234 k--;
emilmont 1:fdd22bb7aa52 235 }
emilmont 1:fdd22bb7aa52 236
emilmont 1:fdd22bb7aa52 237 /* If the count is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 238 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 239 k = count % 0x4u;
emilmont 1:fdd22bb7aa52 240
emilmont 1:fdd22bb7aa52 241 while(k > 0u)
emilmont 1:fdd22bb7aa52 242 {
emilmont 1:fdd22bb7aa52 243 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 244 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 245
emilmont 1:fdd22bb7aa52 246 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 247 k--;
emilmont 1:fdd22bb7aa52 248 }
emilmont 1:fdd22bb7aa52 249
emilmont 1:fdd22bb7aa52 250 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 251 *pOut++ = sum;
emilmont 1:fdd22bb7aa52 252
emilmont 1:fdd22bb7aa52 253 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 254 py = ++pSrc1;
emilmont 1:fdd22bb7aa52 255 px = pIn1;
emilmont 1:fdd22bb7aa52 256
emilmont 1:fdd22bb7aa52 257 /* Increment the MAC count */
emilmont 1:fdd22bb7aa52 258 count++;
emilmont 1:fdd22bb7aa52 259
emilmont 1:fdd22bb7aa52 260 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 261 blockSize1--;
emilmont 1:fdd22bb7aa52 262 }
emilmont 1:fdd22bb7aa52 263
emilmont 1:fdd22bb7aa52 264 /* --------------------------
emilmont 1:fdd22bb7aa52 265 * Initializations of stage2
emilmont 1:fdd22bb7aa52 266 * ------------------------*/
emilmont 1:fdd22bb7aa52 267
emilmont 1:fdd22bb7aa52 268 /* sum = x[0] * y[srcBLen-1] + x[1] * y[srcBLen-2] +...+ x[srcBLen-1] * y[0]
emilmont 1:fdd22bb7aa52 269 * sum = x[1] * y[srcBLen-1] + x[2] * y[srcBLen-2] +...+ x[srcBLen] * y[0]
emilmont 1:fdd22bb7aa52 270 * ....
emilmont 1:fdd22bb7aa52 271 * sum = x[srcALen-srcBLen-2] * y[srcBLen-1] + x[srcALen] * y[srcBLen-2] +...+ x[srcALen-1] * y[0]
emilmont 1:fdd22bb7aa52 272 */
emilmont 1:fdd22bb7aa52 273
emilmont 1:fdd22bb7aa52 274 /* Working pointer of inputA */
emilmont 1:fdd22bb7aa52 275 px = pIn1;
emilmont 1:fdd22bb7aa52 276
emilmont 1:fdd22bb7aa52 277 /* Working pointer of inputB */
emilmont 1:fdd22bb7aa52 278 pSrc2 = pIn2 + (srcBLen - 1u);
emilmont 1:fdd22bb7aa52 279 py = pSrc2;
emilmont 1:fdd22bb7aa52 280
emilmont 1:fdd22bb7aa52 281 /* count is index by which the pointer pIn1 to be incremented */
emilmont 1:fdd22bb7aa52 282 count = 0u;
emilmont 1:fdd22bb7aa52 283
emilmont 1:fdd22bb7aa52 284 /* -------------------
emilmont 1:fdd22bb7aa52 285 * Stage2 process
emilmont 1:fdd22bb7aa52 286 * ------------------*/
emilmont 1:fdd22bb7aa52 287
emilmont 1:fdd22bb7aa52 288 /* Stage2 depends on srcBLen as in this stage srcBLen number of MACS are performed.
emilmont 1:fdd22bb7aa52 289 * So, to loop unroll over blockSize2,
emilmont 1:fdd22bb7aa52 290 * srcBLen should be greater than or equal to 4 */
emilmont 1:fdd22bb7aa52 291 if(srcBLen >= 4u)
emilmont 1:fdd22bb7aa52 292 {
emilmont 1:fdd22bb7aa52 293 /* Loop unroll over blockSize2, by 4 */
emilmont 1:fdd22bb7aa52 294 blkCnt = ((uint32_t) blockSize2 >> 2u);
emilmont 1:fdd22bb7aa52 295
emilmont 1:fdd22bb7aa52 296 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 297 {
emilmont 1:fdd22bb7aa52 298 /* Set all accumulators to zero */
emilmont 1:fdd22bb7aa52 299 acc0 = 0.0f;
emilmont 1:fdd22bb7aa52 300 acc1 = 0.0f;
emilmont 1:fdd22bb7aa52 301 acc2 = 0.0f;
emilmont 1:fdd22bb7aa52 302 acc3 = 0.0f;
emilmont 1:fdd22bb7aa52 303
emilmont 1:fdd22bb7aa52 304 /* read x[0], x[1], x[2] samples */
emilmont 1:fdd22bb7aa52 305 x0 = *(px++);
emilmont 1:fdd22bb7aa52 306 x1 = *(px++);
emilmont 1:fdd22bb7aa52 307 x2 = *(px++);
emilmont 1:fdd22bb7aa52 308
emilmont 1:fdd22bb7aa52 309 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 310 k = srcBLen >> 2u;
emilmont 1:fdd22bb7aa52 311
emilmont 1:fdd22bb7aa52 312 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 313 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 314 do
emilmont 1:fdd22bb7aa52 315 {
emilmont 1:fdd22bb7aa52 316 /* Read y[srcBLen - 1] sample */
emilmont 1:fdd22bb7aa52 317 c0 = *(py--);
emilmont 1:fdd22bb7aa52 318
emilmont 1:fdd22bb7aa52 319 /* Read x[3] sample */
emilmont 1:fdd22bb7aa52 320 x3 = *(px++);
emilmont 1:fdd22bb7aa52 321
emilmont 1:fdd22bb7aa52 322 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 323 /* acc0 += x[0] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 324 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 325
emilmont 1:fdd22bb7aa52 326 /* acc1 += x[1] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 327 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 328
emilmont 1:fdd22bb7aa52 329 /* acc2 += x[2] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 330 acc2 += x2 * c0;
emilmont 1:fdd22bb7aa52 331
emilmont 1:fdd22bb7aa52 332 /* acc3 += x[3] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 333 acc3 += x3 * c0;
emilmont 1:fdd22bb7aa52 334
emilmont 1:fdd22bb7aa52 335 /* Read y[srcBLen - 2] sample */
emilmont 1:fdd22bb7aa52 336 c0 = *(py--);
emilmont 1:fdd22bb7aa52 337
emilmont 1:fdd22bb7aa52 338 /* Read x[4] sample */
emilmont 1:fdd22bb7aa52 339 x0 = *(px++);
emilmont 1:fdd22bb7aa52 340
emilmont 1:fdd22bb7aa52 341 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 342 /* acc0 += x[1] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 343 acc0 += x1 * c0;
emilmont 1:fdd22bb7aa52 344 /* acc1 += x[2] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 345 acc1 += x2 * c0;
emilmont 1:fdd22bb7aa52 346 /* acc2 += x[3] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 347 acc2 += x3 * c0;
emilmont 1:fdd22bb7aa52 348 /* acc3 += x[4] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 349 acc3 += x0 * c0;
emilmont 1:fdd22bb7aa52 350
emilmont 1:fdd22bb7aa52 351 /* Read y[srcBLen - 3] sample */
emilmont 1:fdd22bb7aa52 352 c0 = *(py--);
emilmont 1:fdd22bb7aa52 353
emilmont 1:fdd22bb7aa52 354 /* Read x[5] sample */
emilmont 1:fdd22bb7aa52 355 x1 = *(px++);
emilmont 1:fdd22bb7aa52 356
emilmont 1:fdd22bb7aa52 357 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 358 /* acc0 += x[2] * y[srcBLen - 3] */
emilmont 1:fdd22bb7aa52 359 acc0 += x2 * c0;
emilmont 1:fdd22bb7aa52 360 /* acc1 += x[3] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 361 acc1 += x3 * c0;
emilmont 1:fdd22bb7aa52 362 /* acc2 += x[4] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 363 acc2 += x0 * c0;
emilmont 1:fdd22bb7aa52 364 /* acc3 += x[5] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 365 acc3 += x1 * c0;
emilmont 1:fdd22bb7aa52 366
emilmont 1:fdd22bb7aa52 367 /* Read y[srcBLen - 4] sample */
emilmont 1:fdd22bb7aa52 368 c0 = *(py--);
emilmont 1:fdd22bb7aa52 369
emilmont 1:fdd22bb7aa52 370 /* Read x[6] sample */
emilmont 1:fdd22bb7aa52 371 x2 = *(px++);
emilmont 1:fdd22bb7aa52 372
emilmont 1:fdd22bb7aa52 373 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 374 /* acc0 += x[3] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 375 acc0 += x3 * c0;
emilmont 1:fdd22bb7aa52 376 /* acc1 += x[4] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 377 acc1 += x0 * c0;
emilmont 1:fdd22bb7aa52 378 /* acc2 += x[5] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 379 acc2 += x1 * c0;
emilmont 1:fdd22bb7aa52 380 /* acc3 += x[6] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 381 acc3 += x2 * c0;
emilmont 1:fdd22bb7aa52 382
emilmont 1:fdd22bb7aa52 383
emilmont 1:fdd22bb7aa52 384 } while(--k);
emilmont 1:fdd22bb7aa52 385
emilmont 1:fdd22bb7aa52 386 /* If the srcBLen is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 387 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 388 k = srcBLen % 0x4u;
emilmont 1:fdd22bb7aa52 389
emilmont 1:fdd22bb7aa52 390 while(k > 0u)
emilmont 1:fdd22bb7aa52 391 {
emilmont 1:fdd22bb7aa52 392 /* Read y[srcBLen - 5] sample */
emilmont 1:fdd22bb7aa52 393 c0 = *(py--);
emilmont 1:fdd22bb7aa52 394
emilmont 1:fdd22bb7aa52 395 /* Read x[7] sample */
emilmont 1:fdd22bb7aa52 396 x3 = *(px++);
emilmont 1:fdd22bb7aa52 397
emilmont 1:fdd22bb7aa52 398 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 399 /* acc0 += x[4] * y[srcBLen - 5] */
emilmont 1:fdd22bb7aa52 400 acc0 += x0 * c0;
emilmont 1:fdd22bb7aa52 401 /* acc1 += x[5] * y[srcBLen - 5] */
emilmont 1:fdd22bb7aa52 402 acc1 += x1 * c0;
emilmont 1:fdd22bb7aa52 403 /* acc2 += x[6] * y[srcBLen - 5] */
emilmont 1:fdd22bb7aa52 404 acc2 += x2 * c0;
emilmont 1:fdd22bb7aa52 405 /* acc3 += x[7] * y[srcBLen - 5] */
emilmont 1:fdd22bb7aa52 406 acc3 += x3 * c0;
emilmont 1:fdd22bb7aa52 407
emilmont 1:fdd22bb7aa52 408 /* Reuse the present samples for the next MAC */
emilmont 1:fdd22bb7aa52 409 x0 = x1;
emilmont 1:fdd22bb7aa52 410 x1 = x2;
emilmont 1:fdd22bb7aa52 411 x2 = x3;
emilmont 1:fdd22bb7aa52 412
emilmont 1:fdd22bb7aa52 413 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 414 k--;
emilmont 1:fdd22bb7aa52 415 }
emilmont 1:fdd22bb7aa52 416
emilmont 1:fdd22bb7aa52 417 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 418 *pOut++ = acc0;
emilmont 1:fdd22bb7aa52 419 *pOut++ = acc1;
emilmont 1:fdd22bb7aa52 420 *pOut++ = acc2;
emilmont 1:fdd22bb7aa52 421 *pOut++ = acc3;
emilmont 1:fdd22bb7aa52 422
emilmont 1:fdd22bb7aa52 423 /* Increment the pointer pIn1 index, count by 1 */
emilmont 1:fdd22bb7aa52 424 count += 4u;
emilmont 1:fdd22bb7aa52 425
emilmont 1:fdd22bb7aa52 426 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 427 px = pIn1 + count;
emilmont 1:fdd22bb7aa52 428 py = pSrc2;
emilmont 1:fdd22bb7aa52 429
emilmont 1:fdd22bb7aa52 430 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 431 blkCnt--;
emilmont 1:fdd22bb7aa52 432 }
emilmont 1:fdd22bb7aa52 433
emilmont 1:fdd22bb7aa52 434 /* If the blockSize2 is not a multiple of 4, compute any remaining output samples here.
emilmont 1:fdd22bb7aa52 435 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 436 blkCnt = (uint32_t) blockSize2 % 0x4u;
emilmont 1:fdd22bb7aa52 437
emilmont 1:fdd22bb7aa52 438 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 439 {
emilmont 1:fdd22bb7aa52 440 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 441 sum = 0.0f;
emilmont 1:fdd22bb7aa52 442
emilmont 1:fdd22bb7aa52 443 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 444 k = srcBLen >> 2u;
emilmont 1:fdd22bb7aa52 445
emilmont 1:fdd22bb7aa52 446 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 447 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 448 while(k > 0u)
emilmont 1:fdd22bb7aa52 449 {
emilmont 1:fdd22bb7aa52 450 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 451 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 452 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 453 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 454 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 455
emilmont 1:fdd22bb7aa52 456 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 457 k--;
emilmont 1:fdd22bb7aa52 458 }
emilmont 1:fdd22bb7aa52 459
emilmont 1:fdd22bb7aa52 460 /* If the srcBLen is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 461 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 462 k = srcBLen % 0x4u;
emilmont 1:fdd22bb7aa52 463
emilmont 1:fdd22bb7aa52 464 while(k > 0u)
emilmont 1:fdd22bb7aa52 465 {
emilmont 1:fdd22bb7aa52 466 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 467 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 468
emilmont 1:fdd22bb7aa52 469 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 470 k--;
emilmont 1:fdd22bb7aa52 471 }
emilmont 1:fdd22bb7aa52 472
emilmont 1:fdd22bb7aa52 473 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 474 *pOut++ = sum;
emilmont 1:fdd22bb7aa52 475
emilmont 1:fdd22bb7aa52 476 /* Increment the MAC count */
emilmont 1:fdd22bb7aa52 477 count++;
emilmont 1:fdd22bb7aa52 478
emilmont 1:fdd22bb7aa52 479 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 480 px = pIn1 + count;
emilmont 1:fdd22bb7aa52 481 py = pSrc2;
emilmont 1:fdd22bb7aa52 482
emilmont 1:fdd22bb7aa52 483 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 484 blkCnt--;
emilmont 1:fdd22bb7aa52 485 }
emilmont 1:fdd22bb7aa52 486 }
emilmont 1:fdd22bb7aa52 487 else
emilmont 1:fdd22bb7aa52 488 {
emilmont 1:fdd22bb7aa52 489 /* If the srcBLen is not a multiple of 4,
emilmont 1:fdd22bb7aa52 490 * the blockSize2 loop cannot be unrolled by 4 */
emilmont 1:fdd22bb7aa52 491 blkCnt = (uint32_t) blockSize2;
emilmont 1:fdd22bb7aa52 492
emilmont 1:fdd22bb7aa52 493 while(blkCnt > 0u)
emilmont 1:fdd22bb7aa52 494 {
emilmont 1:fdd22bb7aa52 495 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 496 sum = 0.0f;
emilmont 1:fdd22bb7aa52 497
emilmont 1:fdd22bb7aa52 498 /* srcBLen number of MACS should be performed */
emilmont 1:fdd22bb7aa52 499 k = srcBLen;
emilmont 1:fdd22bb7aa52 500
emilmont 1:fdd22bb7aa52 501 while(k > 0u)
emilmont 1:fdd22bb7aa52 502 {
emilmont 1:fdd22bb7aa52 503 /* Perform the multiply-accumulate */
emilmont 1:fdd22bb7aa52 504 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 505
emilmont 1:fdd22bb7aa52 506 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 507 k--;
emilmont 1:fdd22bb7aa52 508 }
emilmont 1:fdd22bb7aa52 509
emilmont 1:fdd22bb7aa52 510 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 511 *pOut++ = sum;
emilmont 1:fdd22bb7aa52 512
emilmont 1:fdd22bb7aa52 513 /* Increment the MAC count */
emilmont 1:fdd22bb7aa52 514 count++;
emilmont 1:fdd22bb7aa52 515
emilmont 1:fdd22bb7aa52 516 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 517 px = pIn1 + count;
emilmont 1:fdd22bb7aa52 518 py = pSrc2;
emilmont 1:fdd22bb7aa52 519
emilmont 1:fdd22bb7aa52 520 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 521 blkCnt--;
emilmont 1:fdd22bb7aa52 522 }
emilmont 1:fdd22bb7aa52 523 }
emilmont 1:fdd22bb7aa52 524
emilmont 1:fdd22bb7aa52 525
emilmont 1:fdd22bb7aa52 526 /* --------------------------
emilmont 1:fdd22bb7aa52 527 * Initializations of stage3
emilmont 1:fdd22bb7aa52 528 * -------------------------*/
emilmont 1:fdd22bb7aa52 529
emilmont 1:fdd22bb7aa52 530 /* sum += x[srcALen-srcBLen+1] * y[srcBLen-1] + x[srcALen-srcBLen+2] * y[srcBLen-2] +...+ x[srcALen-1] * y[1]
emilmont 1:fdd22bb7aa52 531 * sum += x[srcALen-srcBLen+2] * y[srcBLen-1] + x[srcALen-srcBLen+3] * y[srcBLen-2] +...+ x[srcALen-1] * y[2]
emilmont 1:fdd22bb7aa52 532 * ....
emilmont 1:fdd22bb7aa52 533 * sum += x[srcALen-2] * y[srcBLen-1] + x[srcALen-1] * y[srcBLen-2]
emilmont 1:fdd22bb7aa52 534 * sum += x[srcALen-1] * y[srcBLen-1]
emilmont 1:fdd22bb7aa52 535 */
emilmont 1:fdd22bb7aa52 536
emilmont 1:fdd22bb7aa52 537 /* In this stage the MAC operations are decreased by 1 for every iteration.
emilmont 1:fdd22bb7aa52 538 The count variable holds the number of MAC operations performed */
emilmont 1:fdd22bb7aa52 539 count = srcBLen - 1u;
emilmont 1:fdd22bb7aa52 540
emilmont 1:fdd22bb7aa52 541 /* Working pointer of inputA */
emilmont 1:fdd22bb7aa52 542 pSrc1 = (pIn1 + srcALen) - (srcBLen - 1u);
emilmont 1:fdd22bb7aa52 543 px = pSrc1;
emilmont 1:fdd22bb7aa52 544
emilmont 1:fdd22bb7aa52 545 /* Working pointer of inputB */
emilmont 1:fdd22bb7aa52 546 pSrc2 = pIn2 + (srcBLen - 1u);
emilmont 1:fdd22bb7aa52 547 py = pSrc2;
emilmont 1:fdd22bb7aa52 548
emilmont 1:fdd22bb7aa52 549 while(blockSize3 > 0)
emilmont 1:fdd22bb7aa52 550 {
emilmont 1:fdd22bb7aa52 551 /* Accumulator is made zero for every iteration */
emilmont 1:fdd22bb7aa52 552 sum = 0.0f;
emilmont 1:fdd22bb7aa52 553
emilmont 1:fdd22bb7aa52 554 /* Apply loop unrolling and compute 4 MACs simultaneously. */
emilmont 1:fdd22bb7aa52 555 k = count >> 2u;
emilmont 1:fdd22bb7aa52 556
emilmont 1:fdd22bb7aa52 557 /* First part of the processing with loop unrolling. Compute 4 MACs at a time.
emilmont 1:fdd22bb7aa52 558 ** a second loop below computes MACs for the remaining 1 to 3 samples. */
emilmont 1:fdd22bb7aa52 559 while(k > 0u)
emilmont 1:fdd22bb7aa52 560 {
emilmont 1:fdd22bb7aa52 561 /* sum += x[srcALen - srcBLen + 1] * y[srcBLen - 1] */
emilmont 1:fdd22bb7aa52 562 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 563
emilmont 1:fdd22bb7aa52 564 /* sum += x[srcALen - srcBLen + 2] * y[srcBLen - 2] */
emilmont 1:fdd22bb7aa52 565 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 566
emilmont 1:fdd22bb7aa52 567 /* sum += x[srcALen - srcBLen + 3] * y[srcBLen - 3] */
emilmont 1:fdd22bb7aa52 568 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 569
emilmont 1:fdd22bb7aa52 570 /* sum += x[srcALen - srcBLen + 4] * y[srcBLen - 4] */
emilmont 1:fdd22bb7aa52 571 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 572
emilmont 1:fdd22bb7aa52 573 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 574 k--;
emilmont 1:fdd22bb7aa52 575 }
emilmont 1:fdd22bb7aa52 576
emilmont 1:fdd22bb7aa52 577 /* If the count is not a multiple of 4, compute any remaining MACs here.
emilmont 1:fdd22bb7aa52 578 ** No loop unrolling is used. */
emilmont 1:fdd22bb7aa52 579 k = count % 0x4u;
emilmont 1:fdd22bb7aa52 580
emilmont 1:fdd22bb7aa52 581 while(k > 0u)
emilmont 1:fdd22bb7aa52 582 {
emilmont 1:fdd22bb7aa52 583 /* Perform the multiply-accumulates */
emilmont 1:fdd22bb7aa52 584 /* sum += x[srcALen-1] * y[srcBLen-1] */
emilmont 1:fdd22bb7aa52 585 sum += *px++ * *py--;
emilmont 1:fdd22bb7aa52 586
emilmont 1:fdd22bb7aa52 587 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 588 k--;
emilmont 1:fdd22bb7aa52 589 }
emilmont 1:fdd22bb7aa52 590
emilmont 1:fdd22bb7aa52 591 /* Store the result in the accumulator in the destination buffer. */
emilmont 1:fdd22bb7aa52 592 *pOut++ = sum;
emilmont 1:fdd22bb7aa52 593
emilmont 1:fdd22bb7aa52 594 /* Update the inputA and inputB pointers for next MAC calculation */
emilmont 1:fdd22bb7aa52 595 px = ++pSrc1;
emilmont 1:fdd22bb7aa52 596 py = pSrc2;
emilmont 1:fdd22bb7aa52 597
emilmont 1:fdd22bb7aa52 598 /* Decrement the MAC count */
emilmont 1:fdd22bb7aa52 599 count--;
emilmont 1:fdd22bb7aa52 600
emilmont 1:fdd22bb7aa52 601 /* Decrement the loop counter */
emilmont 1:fdd22bb7aa52 602 blockSize3--;
emilmont 1:fdd22bb7aa52 603
emilmont 1:fdd22bb7aa52 604 }
emilmont 1:fdd22bb7aa52 605
emilmont 1:fdd22bb7aa52 606 /* set status as ARM_MATH_SUCCESS */
emilmont 1:fdd22bb7aa52 607 status = ARM_MATH_SUCCESS;
emilmont 1:fdd22bb7aa52 608 }
emilmont 1:fdd22bb7aa52 609
emilmont 1:fdd22bb7aa52 610 /* Return to application */
emilmont 1:fdd22bb7aa52 611 return (status);
emilmont 1:fdd22bb7aa52 612
emilmont 1:fdd22bb7aa52 613 #else
emilmont 1:fdd22bb7aa52 614
emilmont 1:fdd22bb7aa52 615 /* Run the below code for Cortex-M0 */
emilmont 1:fdd22bb7aa52 616
emilmont 1:fdd22bb7aa52 617 float32_t *pIn1 = pSrcA; /* inputA pointer */
emilmont 1:fdd22bb7aa52 618 float32_t *pIn2 = pSrcB; /* inputB pointer */
emilmont 1:fdd22bb7aa52 619 float32_t sum; /* Accumulator */
emilmont 1:fdd22bb7aa52 620 uint32_t i, j; /* loop counters */
emilmont 1:fdd22bb7aa52 621 arm_status status; /* status of Partial convolution */
emilmont 1:fdd22bb7aa52 622
emilmont 1:fdd22bb7aa52 623 /* Check for range of output samples to be calculated */
emilmont 1:fdd22bb7aa52 624 if((firstIndex + numPoints) > ((srcALen + (srcBLen - 1u))))
emilmont 1:fdd22bb7aa52 625 {
emilmont 1:fdd22bb7aa52 626 /* Set status as ARM_ARGUMENT_ERROR */
emilmont 1:fdd22bb7aa52 627 status = ARM_MATH_ARGUMENT_ERROR;
emilmont 1:fdd22bb7aa52 628 }
emilmont 1:fdd22bb7aa52 629 else
emilmont 1:fdd22bb7aa52 630 {
emilmont 1:fdd22bb7aa52 631 /* Loop to calculate convolution for output length number of values */
emilmont 1:fdd22bb7aa52 632 for (i = firstIndex; i <= (firstIndex + numPoints - 1); i++)
emilmont 1:fdd22bb7aa52 633 {
emilmont 1:fdd22bb7aa52 634 /* Initialize sum with zero to carry on MAC operations */
emilmont 1:fdd22bb7aa52 635 sum = 0.0f;
emilmont 1:fdd22bb7aa52 636
emilmont 1:fdd22bb7aa52 637 /* Loop to perform MAC operations according to convolution equation */
emilmont 1:fdd22bb7aa52 638 for (j = 0u; j <= i; j++)
emilmont 1:fdd22bb7aa52 639 {
emilmont 1:fdd22bb7aa52 640 /* Check the array limitations for inputs */
emilmont 1:fdd22bb7aa52 641 if((((i - j) < srcBLen) && (j < srcALen)))
emilmont 1:fdd22bb7aa52 642 {
emilmont 1:fdd22bb7aa52 643 /* z[i] += x[i-j] * y[j] */
emilmont 1:fdd22bb7aa52 644 sum += pIn1[j] * pIn2[i - j];
emilmont 1:fdd22bb7aa52 645 }
emilmont 1:fdd22bb7aa52 646 }
emilmont 1:fdd22bb7aa52 647 /* Store the output in the destination buffer */
emilmont 1:fdd22bb7aa52 648 pDst[i] = sum;
emilmont 1:fdd22bb7aa52 649 }
emilmont 1:fdd22bb7aa52 650 /* set status as ARM_SUCCESS as there are no argument errors */
emilmont 1:fdd22bb7aa52 651 status = ARM_MATH_SUCCESS;
emilmont 1:fdd22bb7aa52 652 }
emilmont 1:fdd22bb7aa52 653 return (status);
emilmont 1:fdd22bb7aa52 654
emilmont 1:fdd22bb7aa52 655 #endif /* #ifndef ARM_MATH_CM0 */
emilmont 1:fdd22bb7aa52 656
emilmont 1:fdd22bb7aa52 657 }
emilmont 1:fdd22bb7aa52 658
emilmont 1:fdd22bb7aa52 659 /**
emilmont 1:fdd22bb7aa52 660 * @} end of PartialConv group
emilmont 1:fdd22bb7aa52 661 */